Isabel Valera appointed Group Leader of the new independent Probabilistic Learning Group at the Max Planck Institute for Intelligent Systems
The award-winning scientist specializes in research aiming to make machine learning methods more flexible, robust, and fair.
Tübingen – Isabel Valera, who joined the Max Planck Institute for Intelligent Systems (MPI-IS) as a Research Group Leader in Professor Bernhard Schölkopf’s Empirical Inference Department in late 2017, took over the new independent Probabilistic Learning Research Group at MPI-IS on July 1, 2019.
“I am very pleased about this opportunity to continue my career at MPI-IS,” says Valera. “Since I first came to the institute almost two years ago, I have worked with some of the world’s best minds in the field of machine learning. I’m very excited to be building my new group and driving research in the field forward.”
The Probabilistic Learning Group’s research focuses on developing machine learning methods that are flexible, robust, and fair. Flexible means they are capable of modeling complex real-world data, which are often heterogeneous in nature and collected over time. Secondly, Valera’s research aims to improve the robustness of algorithms. An algorithm is considered robust when it is able to point out “what it does not know”. This means that, in addition to making predictions, it also expresses its confidence about the decision. Finally, the group is researching ways to make algorithms fairer – if they are part of important decision-making processes, the outcomes should be fair. To this end, she has worked on translating legal definitions of fairness into metrics that can be quantified using observed data, as well as on designing new algorithms that are not only accurate, but also fair according to a given definition of fairness.
“In my research, I work on a broad range of applications, from medicine and psychiatry to social and communication systems. Recently, we began putting a special focus on consequential decision making in several domains, including hiring processes, pre-trial bail, or loan approval”, Valera explains.
Valera is an accomplished, award-winning scientist. She recently received a DFG Cluster of Excellence grant from the “Machine Learning: New Perspectives for Science” cluster at the University of Tübingen, which provides her with research funding over a period of seven years. She also held a Max Planck Society Minerva Fast-Track Fellowship, which aims to promote career advancement for women in science. Prior to this, Valera received a Humboldt Post-Doctoral Fellowship and completed her Ph.D. in Multimedia and Communications at the Universidad Carlos III de Madrid (Spain).
Over the course of her academic career to date, Isabel Valera has been nominated for and won several awards, among them the 2017 Best Paper Award Honorable Mention at theInternational World Wide Web Conference (WWW), one of the world’s leading data science conferences, for her paper “Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment.”
Next year, Valera will act as Program Chair of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD 2020) in Ghent, Belgium.